Women’s average scores on the SAT have always been lower than men’s—even though they receive higher average grades in all courses in high school and college. The SAT gender gap was first reported in Ms. in 1985; although the difference in average scores between the sexes has narrowed since then from 59 to 20 points, it has never been eliminated. 53 percent of SAT test takers are women, but the gate-keeping college entrance exam is underestimating their ability to succeed.
Last year, Educational Testing Service (ETS), which publishes the SAT, had a chance to even the score between the sexes even further with a major revision of its test. It has chosen, instead, to widen the gender gap for the highest-performing women.
According to an in-depth analysis by Art Sawyer—one of the most knowledgeable leaders in test preparation—the proportion of female students in top-scoring ranges has declined in comparison to the older SAT. Average SAT score differences between all males and females on the new SAT fell to 20 points, which is better than the 24 points of the previous SAT, but this gender gap increases to 40 points when comparing male and female students with the highest scores. According to Sawyer, 45 percent more men than women are in the top score range of 1400-1600. On the old test, the gender gap at the highest score range was 31 percent. Because the College Board stopped reporting detailed information for the test sections by gender and race in 2015, it is difficult to say how this new gender gap has been created.
ETS has not changed the subject matter on the test as much as it has shifted the scoring. The previous reading and writing sections each counted for 33 percent of a person’s score; they have now been combined into one section called Evidence-Based Reading and Writing which counts for 50 percent. Women’s reading scores were often slightly lower than men’s, but on the writing section they were always higher, sometimes by 14 points—which would offset their historically lower math scores and narrow the gap. The new test has thus decreased the weighting of a subject where women were excelling—and now, the section where the gender gap is consistently the largest counts for half of the total score. (Even more significant is the removal of the essay question, where women outperformed men in the old writing section. Female writing scores have now declined until they are similar to men’s.)
Though the new test was touted as a closer reflection of math taught in the classroom, where women receive higher average grades than men, this doesn’t seem to be the case. ETS didn’t just change the weighting of the math section—according to Sawyer, it “dramatically overhauled the exam,” reducing geometry questions and shifting the focus to algebra. There are many different types of math ability, and changes in the math tested can have a significant impact on how well students perform. Earlier research, by myself and others, has shown that the gender gap on the math section could be reduced or even eliminated if there was a fair balance of questions that favor one sex or the other.
ETS has always known which test questions favor women and which favor men. A 30-minute ungraded section is included in every SAT to identify how hard new questions are for men and women in all ethnic groups before they are added to the test. Test writers then decide whether to include them in future tests. In 1989, Sociologist Jim Loewen and I found 17 questions that were considerably easier for one sex than the other; seven of them were in the verbal section, with four favoring women and three favoring men, and 10 questions in the math section all favored men. Our research convinced us that ETS intentionally creates a test that favors men, especially at the highest score levels—where cut scores are vitally important to future success.
As a result of this new gender gap, fewer high-performing women are receiving the highest scores required for the most selective colleges. They are also receiving fewer of the scholarships given to high scoring students and are denied entrance into college honors classes which use test scores. Lower scores also have a negative impact on women’s self esteem, causing them to limit their opportunities and lower their expectations for the rest of their lives.
Before the new SAT was written, The College Board Research Report of 2007 discussed the upcoming revision, stating that subgroup performances in the new test would not be exacerbated. “If they moved at all they would be in the direction of reducing the differences,” the organization explained. “They will be compared to historical subgroup data to ensure that the gaps that exist on the current test do not widen.” This worthy goal seems to have been ignored.